Professional Analog Mastering.
Professional Analog Mastering.
226K+ subscribers
First, it’s going to stay simple to avoid unwanted plugin artifacts and keep the relationships between each processing stage straightforward.
Second, I’ll introduce as little undesirable distortion as possible - so I’m talking about aliasing, pre-ringing, intermodulation distortion, noise, and so on - basically, any distortion type that doesn’t blend in naturally with the music like harmonic distortion does.
Third, I’ll utilize well-designed processors that I’ve tested and know don’t introduce anything I don’t want. That said, feel free to substitute what I’m using with stock plugins if needed.
Lastly, I’m mastering at a 96kHz sampling rate to reduce the need for oversampling. Some will still be needed, but much less so than in a 48kHz session or lower. Since the chain is so straightforward, this shouldn’t cause any issues with the CPU.
Side note: if you’re a member and you want to follow along, you can download this track from the multi-tracks for mixing practice tab. The track is titled Borderline, and I’m using the stereo mix that’s included in the download for this video.
Let’s listen to the before-and-after. I won’t use gain compensation for this first demo, but I will when we discuss limiting later on.
Watch the video to learn more >
I’m using zero latency because, although subtle, the pre-ringing distortion introduced by linear phase filters blurs the transients of the lows, which I want to avoid.
As we introduce this EQ, notice how subtle the changes are. The first filters I’d recommend trying are a dip in the mid-image around 350Hz and a boost to the side image one-half step above the mid-image filter. Additionally, the mid-image dip should be slightly more aggressive than the side-image boost.
This maintains the range's balance but alters the stereo width to reduce muddiness and clashing in the region. The slight offset in frequencies subtly expands the region as well through mild phase interference.
For this track in particular, there’s a strong bass fundamental that needed to be adjusted - and some nasal tones from the vocal that were too prominent. By using M/S filters to target the offending ranges more accurately, I’m reducing the need for additional filters or subsequent processors to fix any overcorrections that would occur with regular stereo filters.
A gentle 6dB/Octave high pass on the side image will center the lows, and very mildly center the low mids - creating a clean and focused sound without changing too much.
A subtle shelf boosts the kick in the mid-image, while a high shelf on the side image creates clean-sounding air and width.
Let’s take a listen.
Watch the video to learn more >
The cleanest compressor I’ve found is the Pro-C 2, set to its mastering algorithm. It introduces little distortion when attenuating the signal, which makes it perfect for very subtly attenuating peaks and amplifying quieter details.
Notice I’m only achieving half a dB of compression - using a high threshold with a somewhat softer knee, a low ratio, and a quick automatic release.
This compression isn’t for altering the track’s timbre - in fact, that’s the opposite of what I want to do. Instead, it reduces the dynamic range without any noticeable change.
Let’s listen and notice how it’s nearly impossible to hear a difference between the before and after.
Watch the video to learn more >
Usually, saturation is used to both fill the lows and mids, while also introducing harmonics to the highs to brighten the sound - but since this second part would introduce aliasing distortion and phase cancellation in the highs, I have an alternative I’ll show over the next 2 chapters.
First, I’ll introduce frequency-specific saturation and isolate the effect to right above the snare drum’s fundamental. This means I’ll generate harmonics from the kick, bass, and snare.
Notice that the band on the higher frequency range has been turned off completely.
Next, I’ll saturate the low band, but only when a transient is present, by setting up an envelope follower and linking it to the drive dial. This ensures that any harmonics are tied to transients, making it much less noticeable if a generated harmonic frequency is not in-key with the song.
Additionally, I’ll time the release of the distortion to the BPM of the song, at least to the best of my ability.
A clean tube setting works well for this track, but experiment to find what works best for your song.
Lastly, I’ll reduce the overall amplitude of the low-frequency band to compensate for the added amplitude of the harmonics.
Doing this causes the majority of the harmonic’s amplification to occur above the cut-off frequency or in the mids - which will help avoid muddiness due to 2nd or 3rd-order harmonics.
Let’s listen.
Watch the video to learn more >
If I’m determined to keep a master as distortion-free as possible but want a bright and airy sound, this is a great alternative to high-hz. saturation or exciters.
First, I’ll create a send or bus from the stereo mix. On the auxiliary track, I’ll insert the free plugin MSED and mute the mid-image, leaving only the sides.
Then, with a linear-phase EQ, I’ll attenuate the lows and isolate the high frequencies, as well as boost the air with a shelf. Although I was avoiding linear phase filters earlier, pre-ringing is much less noticeable in high frequencies - plus, we’re about to compress this range and amplify quieter details which will help mask some of that pre-ringing.
Using the Pro-C2 set the the mastering algorithm, I’ll use a soft knee and low ratio to gently compress the isolated side-image highs. With the range function, I’ll keep the attenuation to 7.5dB at most to ensure that the compression doesn’t become audible. Lastly, I’ll make the release program dependent and enable auto-make-up gain to amplify the quieter details.
At this point, I can adjust the aux track’s channel fader to control the track’s brightness and the stereo width of these dense high frequencies.
Let’s listen to how this effect creates a clean, bright, and open sound for the track.
I’ve discussed this limiter before, but Elephant by Voxengo is as close as I’ve found to transparent limiting.
With it, I’ll achieve up to 3.5dB of attenuation, and with the E-4 algorithm, I won’t be able to hear much of a difference in the timbre.
2x oversampling is helpful at this point—it couldn’t be completely avoided, but using only 1 instance of it, and at a low setting, still results in a clean-sounding master.
Let’s listen to the A/B without gain compensation so you can hear how it alters the track with loudness changes included. Then, we’ll do the A/B with amplitude compensation and notice how the track’s timbre is only subtly altered.
Last thing, by the end of this chain, we achieved a LUFS of -11, which is somewhat quiet but more than enough considering normalization will alter its loudness.